Report No. 44109-AL Albania Local Finance Policy Note Programmatic Public Expenditure and Institutional Review September 2, 2008 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank Albania Local Government Finance Policy Note EXECUTIVESUMMARY............................................................................................................................I OVERVIEW LOCALGOVERNMENTFINANCE OF .................. .. ............................................................ 1 Background ........................................................................................................................................ 1 Functions ...................... ....................................................................................................................2 Social Sectors ....................................................................................................................................... 2 Urban Housekeeping and Infrastructure Services ............................................................................ 3 Administrative Functions ........................ ...... ............................................................................... 4 Revenue Assignment ......................... ... .........................................................................................4 Do LOCALGOVERNMENTSHAVEENOUGHMONEY?....................... .. ................................................ 5 CHANGINGTAXASSIGNMENTS .......................... ............................................................................. 7 Small Business Tax ............................................................................................................................ 7 Property Tax.................................................................................................................................. 8 Personal Lncome Tax.......................................................................................................................... I I FINE-TUNING TRANSFERSYSTEM THE ...................... .. ...................................................................... 12 Stability ...................... ............................................................................................................... .. 12 Equalization........................ ... ................................................................................................... 13 Earmarking ......................... ...........................................................................................................15 FURTHERDECENTRALIZATION ........................................................................................................... 16 Water Company Decentralization ....................................................................................................16 Managing Teachers' Salaries ........................... .................................................................................. 18 FINANCINGMUNICIPALCAPITALINVESTMENT ................................................................................. 18 Competitive Grants........................................................................................................................ 18 General Municipal Infrastructure............................................................................................. 2 0 Water Supply............... ... ........................................................................................................ 22 Education and Health .................................................................................................................... 24 Municipal Borrowing ......................... .. ....................................................................................... 25 Ceilings on Local Government Borrowing ................................................................................ 25 Macroeconomic Limitations on Aggregate Debt ................................................................... -26 Fiscal Distress ..................... .................................................................................................... .. 27 CHANGES THE STRUCTURE LOCAL GOVERNMENT IN OF .................................................................. -28 Regional Governments................................................................................................................. 28 . Consolidation................................................................................................................................. 29 Table 1: Average Population of Municipalities in Comparator European Countries ................................. 2 Table 2: Sources of Local Government Discretionary Revenues (2007).................................................. 5 Table 3: Indicators of Variations in Per Capita Discretionary Revenues................................................. 14 Table 4: Current Operating Subsidies to Water Companies (LEK mn).................................................. 17 Table 5: Budget Allocations for Competitive Grants.................. ... ...................................................... 20 Table 6: Performance Indicators (measured as percentage change over previous period) ....................... 23 Table 7: Need Indicators (measured as absolute level)............................................................................ 23 Figures Figure 1: Trends in Local Expenditures....................................................................................................... 3 Figure 2: Trends in Level and Composition of Revenues.......................................................................... 4 Figure 3: Local Expenditure as % GDP European Comparators .............................................................. 7 Figure 4: Property Tax Revenues European Comparators ....................................................................... 10 Figure 5: Local Personal Income Tax Revenues European Comparators.............................. .. ..............11 Figure 6:Trends in Unconditional Grants................. ... ........................................................................ 12 Figure 7: Variation in Per Capita Revenues......................................................................................... 14 Box 1: Unconditional Grant Distribution Formula (2008)........................................................................ 6 Box 2: Juggling the Costs and Standards of Property Tax Administration ...................... ...........................10 Box 3: Calculating the Operating Subsidy for Water Companies................ ...............................................17 Box 4: ADF Criteria for Evaluating Municipal Infrastructure Projects ................... .... ..................... 21 Box 5: Controlling Arrears..................... .. ..............................................................................................28 Acknowledgements This Note was prepared by William Dillinger (ECSPE). with assistance and guidance from Alia Moubayed. Erjon Luci. Michael Webster. Arben Bakllamaja. Richard Martin Humphreys. Yan Zhang and Sabina Ymeri.Management oversight was provided by Bernard Funck and Ronald E. Myers (ECSPE). EXECUTIVE SUMMARY Albania has undertaken major reforms in its system of local government finance since 2000. What had been a system in which local functions were ambiguous and financing was largely provided through tightly controlled earmarked grants is now one in which functions are relatively clear and local governments have more autonomy over the allocation of funds. A new system of competitive grants for infrastructure investment has been introduced. Parliament has enacted a new law on local borrowing. Within this framework, however, several controversies remain which are addressed in this policy note. Should the overall level of local resources be increased? There is not a compelling case for increasing local governments' share of public resources. Although local government spending, as a percent of GDP, is the lowest of any comparable country in Europe, local expenditure responsibilities are correspondingly limited. Unlike in many other European countries, Albanian local governments do not finance teachers' salaries or contribute to health care costs. Spending on social assistance is largely financed out of earmarked grants. Once these differences in expenditure responsibilities are considered, Albanian local government resources are not out of line with European norms. In the absence of any other evidence that local functions are under-funded--relative to those of the central government--there is no compelling case for increasing the local share. Should local governments be assigned different, more productive tax instruments? Albanian local governments are unusual in that the principal source of local taxes is a tax on small business. This is not a particularly cost effective means of raising revenue, as the volume of taxable activity per taxpayer is inherently limited. The Government is now considering ways to increase the yield of one of the other taxes assigned to local government: the building tax. Efforts to improve coverage and shift to market based valuations may pay off, but only if technical standards are appropriately modest and equal effort is paid to raising the nominal rate and improving collection enforcement. Other European countries allow local governments to retain a share of the personal income tax collected within their boundaries. This is a long term option for Albania but not a short term one. At present the personal income tax base is so concentrated in Tirana that it would yield little revenue in most jurisdictions. Should the rules governing the allocation of intergovernmental transfers be changed? Local governments derive about half their revenues from unconditional grants. The core criteria for distributing the grant (total population, urban population, and land area) are reasonably transparent and equitable. Additions to the transfer formula have made it increasingly convoluted over time, however. Although the transfer system leaves substantial variations in per capita discretionary revenues, the case for further equalization is not strong. Equalization is desirable where local governments finance functions with major distributional implications, such as education, health, or social assistance. One would not want spending in these functions to depend on the strength of the local tax base. But in Albania, these functions are financed--either directly or through earmarked grants-by the central government. Albanian local governments instead provide economic services: principally the maintenance of streets and rural roads and various urban 'housekeeping' functions. Under these circumstances, a system that leaves resources in the jurisdictions where they are generated makes sense. In the absence of any major redistributional functions to be financed from discretionary revenues, the case for leaving the current level of equalization as it is is compelling. Should the rules governing the allocation of intergovernmental transfers be codified in a permanent form of legislation? The principal shortcoming of the current transfer system is its unpredictability. The overall level of the transfer is subject to annual budget negotiations. The formula for distributing the transfer changes from year to year. The result is a high level of uncertainty and vulnerability to short term-political considerations. The Government should consider enshrining the rules governing unconditional transfers in a more permanent form of legislation, such as the local government organic law or the organic law on budget preparation. Such legislation should establish a mechanism for determining the overall amount of the transfer each year-as a fixed share of aggregate national tax revenue, for example. The legislation should also specify the division of transfers between classes of jurisdictions and the criteria to be used to distribute funds to each jurisdiction within each class. To provide for flexibility, these provisions could be subject to periodic review at, say, five year intervals, by a high level group including representatives of the relevant ministries at the central level and the associations of local governments. How can local governments cope with the take-over of loss-making water companies? The ongoing decentralization of water utilities could potentially saddle local governments with significant financial obligations. The Government is committed to phasing out recurrent subsidies to the water companies in the next few years. Such subsidies would then have to be paid by their new local government owners. Water utilities can reduce their need for subsidies in a variety of ways: expanding metering (to reduce consumption and therefore production costs), shedding staff, and reducing unaccounted for water (through leak detection and registration of illegal connections). The Government should encourage the utilities to take such measures even as decentralization proceeds by imposing a fixed, uniform ceiling on the level of operating subsidies it provides to each utility and gradually reducing the level of the subsidy over time. How can the allocation of competitive grants be improved? Competitive grants are currently used to finance small capital investments in four sectors: (1) local roads and other municipal infrastructure, (2) water supply, (3) education, and (4) health'. Separate grant programs exist for each sector. While the current evaluation process is successful in winnowing technically sound projects out of hundreds of applications, each program has certain weaknesses. In the case of the municipal infrastructure grant, the problem is a tendency to approve more projects than can be fully funded in the current year. As a result, projects are only partially funded. This results in costly delays and interruptions. In the short term, this could be solved by refusing all new applications until the existing backlog is completed. In the long term, it could be addressed by tightening the evaluation criteria in order to reduce the number of approved projects and fully funding those that are approved. In education, the problem is consistency with national investment priorities. At present, competitive grants are geographically dispersed. The distribution of grants tends to reflect the geographical distribution of applications, along with their relative technical quality and de facto caps on individual project size. But in Albania, geographical priorities vary. As a result of rapid migration, some local governments have too many classrooms; others too few. Although the MOE has defined a set of geographical investment priorities, it must rely on requests from local governments to implement them. This problem could be addressed by imposing an overriding 'geographical priorities' criterion on the allocation of the grants or by simply re-centralizing investment decisions in the Ministry of Education. Major revisions to the competitive grant for water are currently under consideration. In broad terms, they would add explicit performance and needs criteria into the allocation rules. The specific formula now under consideration is commendable, although it will require some refinements. The Government will also have to determine whether to apply the formula to larger scale water investments, including those financed by donors. Are the recently-introduced controls on local government borrowing appropriate? The new local government borrowing law imposes very appropriate quantitative controls on the volume of local h he competitivegrant for health was not evaluated for this report. borrowing. The law also requires aggregate local government debt to fit within an aggregate debt ceiling for the public sector as a whole. This is defined as 60 percent of GDP. Additional regulations will eventually be required to specify how the local government share of the aggregate public borrowing limit will be determined and how that amount will be allocated among individual jurisdictions. For the short term, the issue appears to be moot. This is, first, because the three quantitative restrictions will limit total local government debt to about 3 percent of GDP. And second, because, in the absence of central government guarantees, the private financial sector is likely to limit its exposure to long term local government debt. What long-term changes in the organization of local governments are needed? In the long term Albania must also grapple with certain organizational issues. One is the role of regional governments, which now lack any major clearly defined function. While the Government's decentralization strategy proposes a range of possible functions for regional governments, most of these roles could be carried out through other means. The other organizational issue is the consolidation of small communes. The Government's decentralization strategy proposes the eventual consolidation of small local governments into larger units. International experience, however, suggests that the efficiency gains of consolidation are limited and may not be sufficient to justify the political costs of doing so. Efforts to consolidate local governments should consider alternative solutions-such as inter-municipal agreements and joint service companies-before proceeding. 1. Albania has undertaken major reforms in its system of local government finance since 2000. What had been a system in which local functions were ambiguous and financing was largely provided through tightly controlled earmarked grants is now one in which functions are relatively clear and local governments have more autonomy over the allocation of funds for functions that lie within their exclusive responsibility. A new system of competitive grants for infrastructure investment has been introduced. Parliament has enacted a new law on local borrowing. The Government has also issued a new decentralization strategy, which lays out general directions for reform. 2. Still, several controversies remain to be addressed. In particular: Should the overall level of local revenues be increased? Should local governments be assigned different, more productive tax instruments? Should the rules governing the allocation of intergovernmental transfers be simplified? Should they be made more equalizing? Should they be codified in a permanent form of legislation? How can local governments cope with the take-over of loss-making water companies? 3. The system of financing for local capital investment involves an additional set of issues. These include: How to improve the allocation of competitive grants, and How to implement the recently introduced controls on municipal borrowing. 4. In the longer term, Albania must also grapple with organizational issues, particularly the role of regional governments and the possible merits of consolidating small local governments. 5 . This note addresses these issues. Background 6. The current (1998) Constitution defines communes and municipalities (hereafter, local governments) as the basic units of local government. Local governments are legal entities and 'perform all the duties of self-government, with the exception of those that the law gives to other units of local government'. There are a total of 374 local government units, consisting of 65 municipalities and 309 communes. While municipalities govern urban areas, and communes, rural areas, there is no substantive legal distinction between them. The average size of local governments (in terms of population) is about 8,400. As shown in Table 1, this falls within the mid-range for European countries, and is comparable to the figures for several neighboring countries, including Croatia, Greece, and ltaly.* he Constitution recognizes a second tier of subnational government, termed the 'region'. As defined in the Constitution, regions consist of several 'basic units of local government with traditional, economic, and social ties and joint interests.' At present, regional governments function largely as associations of local governments. They have no directly elected officials, few functions, and limited budgets. 7. Overall, local government expenditures have increased over the last decade, both in nominal terms Table 1: Average Population of and as a proportion of total general government Municipalities in Comparator European expenditure. As shown in Figure 1, local expenditures Countries increased by nearly 60 percent between 1998 and 2004, Country Avg. Pop. (000) before stabilizing at their current levels. As a percentage Lithuania 66.3 of total government expenditures, local spending has Sweden Bulgaria declined somewhat since 2004, due to the rapid growth of Netherlands central government expenditures. Macedonia Denmark Functions Serbia Belgium Poland 15.5 8. Albanian local governments perform several Finland 10.9 distinct categories of functions. While the social sectors Greece 10.7 consume the largest share of local budgets, local Noway 9.0 management responsibilities in this area are limited. Most Albania 8.4 of local government discretionary spending is devoted to Croatia 8.0 expenditures on streets and rural roads, various urban Italy 7.1 'housekeeping' functions, and administrative costs. Romania 6.9 Estonia 5.7 Germany 5.6 Social Sectors Latvia 4.4 Spain 4.9 9. Social assistance. Social assistance is the largest Austria 3.4 single category of local government expenditure. This 3.2 largely consists of a program of direct cash payments to 2.5 Czech Rep. 1.7 low income households (the ndihma ekonomike). It is France 1.6 financed entirely by earmarked grants from the central Source: IMF Government Finance Statistics government and is distributed according to guidelines Yearbook 2007 and country files,Excludes issued by the Ministry of Labor and Social Protection. In intermediatetiers ofsub nationalgovt. principle, funds earmarked for this purpose are allocated according to the number of eligible households in each jurisdiction, on the basis of centrally defined criteria. In practice, local governments have been known to exercise some discretion in the allocation of these funds. 10. Primary and secondary education are, under the municipal organic law, -a shared function of central and local government. In practice, most of the burden of managing and financing this function is borne by the Ministry of Education. Teachers are employees of the Ministry of Education (MOE) which funds the salaries and benefits of individual teachem3 Local governments are, however, responsible for classroom construction, which is funded through conditional grants to local governments. Local governments are also responsible for the operating costs of school buildings--heating, lighting, and building maintenance, which are financed from discretionary revenues. 3~ntil2008, local governments were responsible to distributingGovernment-financed salariesto individual teachers. As of 2008, funding for teachers bypasses local governments entirely. Salaries are instead paid directly to teachers by district offices of the MOE. Note that, following Albanian practice, the funding of teachers' salaries in previous years is not included in municipalrevenue or expendituretables used in this report. 11. Other social services Local governments also spend relatively minor sums on social, cultural and recreational services. These include the Figure Trends in Expenditures organization and management of sports , I institutions and the administration of institutions such as day care centers, elderly homes and orphanages. The role of local governments in health care is minor. At present, the Ministry of Health owns and operates the entire system. The Health Insurance Institute (HII) funds all primary care, including salaries, pharmaceuticals, utilities, and building maintenan~e.~ Urban Housekeeping and Infrastructure Services 12. Local governments have considerably more discretion in the provision of urban housekeeping and infrastructure services. Unlike social protection and teachers salaries, the recurrent costs of these functions are largely financed out of discretionary revenues, which local governments are free to allocate as they see fit. 13. Roads--including urban streets and sidewalks--are the largest single item of spending in this category. Prior to 2003, most road construction and maintenance was a central government responsibility. Local spending on roads, if any, was financed through conditional grants. In 2003, rural, communal, and regional roads were transferred to the local governments. Spending on the construction and maintenance of roads and urban streets is now financed from local discretionary revenues, supplemented by competitive grants. Solid waste management, street cleaning, and park maintenance also consume a significant proportion of local government revenues. Tirana, for example, spends about ten percent of its discretionary budget on solid waste management and street cleaning and another ten percent on parks.5 14. Responsibility for water supply is still in a process of transition. The local government organic law defines water supply as an exclusive function of local government. Many communes, in fact, have long operated free-standing water supply systems in villages. Until late 2007, however, most systems serving larger urban areas were owned by the Ministry of Economy and supervised by the Ministry of Public Works' General Directorate for Water and Sanitation (GDWS). It provided operating subsidies and grants for capital investment directly to all of them. The Government is now in the process of transferring the utilities to local governments (or consortia of local governments in the case of companies serving 4 ~2004, Council of Ministers adopted a Decision on Decentralization Policies for primary health care and public n health protection but it essentially involves a reorganization of management and financing within the national health care system;not a transfer of responsibility to local governments. ' ~ a t aon local spending on urban services is difficult to measure due to the way it is categorized. The Albanian treasury classifies functional expenditures on the basis of the IMF's Government Finance Statistics (GFS) methodology.Functions such as solid waste management,street cleaning, and park maintenance fall under the GFS' 'housing and community services category' but this category also includes the construction, rehabilitation and maintenance of the housing stock (whether publicly or privately owned), urban planning, and street lighting. (Water supply, but not sewerage or solid waste collection, is also included in the GFS definition of housing and community development but is treated separately in the Albanian data.) According to the Albanian classification system, 'housing and community services' accounted for 16 percent of local discretionary spending in 2006. As Albania's housing stock is largely privatized, the majority of this was presumably spent these other urban housekeeping functions. more than one jurisdiction). The Government also intends to phase out operating subsidies to the companies. As discussed later in this report, this could have major financial implications for local governments. Administrative Functions 15. Local governments also carry out a variety of administrative functions. The scale and nature of expenditures in this category cannot be determined with the data at hand. In GFS parlance, such expenditures fall under the rubric of 'general public services'. This category is intended to include only the functions of the chief executive and the legislative branches, including personnel management, financial administration, tax administration and the payment of debt service. In 2006, 'general public services' consumed 37 percent of local government discretionary spending. This may be an overstatement. In Albania, smaller local governments are said to report a wide range of expenditures under this category, including their entire wage bills. Revenue Assignment 16. Local governments derive their revenues from a variety of sources. Roughly half of local government revenues come in the form of conditional (earmarked) grants. As noted earlier, this includes funding for social protection: the ndihma ekonomike. But the remainder of local revenues come with no strings attached and can be allocated by local governments as they wish. As shown in Figure 2, discretionary revenues increased dramatically between 2000 and 2006. This is partly due to growth in unconditional transfers but also due to changes in legislation, which have increased the taxing powers of local governments. 17. Local Taxes and Fees Forty-three Figure 2: Trends in Level and Composition of percent of discretionary revenues are derived Revenues from locally administered taxes and fees. The current structure of local taxes and fees reflects major reforms that went into effect in 45000 El 2003 and 2007. Prior to 2003, local 40000 .CondnlonalGrants government taxes consisted of a one percent 35000 .Uncond~t~onalGrants turnover tax on restaurants, hotels, and bars, 30000 o .Shared T a m separate taxes on hotels, business signs, E 25000 . O ~ ~ ~ ~ L O C ~ I N O ~ - t a x construction, and the use of public space and 20000 -LocalTams and Faes a 60 percent share of the property tax, 15000 distributed on derivation basis. Local 10000 governments were also permitted to impose 5000 0 fees for business licenses and various 8I s g g g g f i W g g - municipal services, including garbage collection. Proceeds from the centrally administered small business tax were used to finance unconditional transfers to local governments but were not legally designated for this purpose. 18. In 2002, Parliament approved a package of legislation (Laws 8982, 8979, 8978). The one percent turnover tax was eliminated and replaced by a profits tax and a tax on small businesses. These were later combined into a single tax on small business, which, since January 2007, has been administered by local governments. 19. Local governments-and particularly Tirana-also derive a significant share of revenue from a construction fee. This is imposed prior to the issuance of a building permit. In Tirana, the rate is 4 percent of the cost of construction, which is based on a notional cost of 300 per square meter. Property taxes make a relatively small contribution to local revenues. Although Parliament increased the local share of the property tax to 100 percent in 2003, the tax contributed only five percent to discretionary revenues in 2007. 20. Unconditional Transfers. The other principal source of local discretionary revenues is unconditional transfers. As shown in Table 2, these contributed nearly half of the total in 2007. 21. Albania's system of unconditional Table 2: Sources of Local Government Discretionary transfers took effect in 2002. Under Law Revenues (2007) 887412001. former block -grants and Revenues certain conditional grants (including Source (LK mn) Percent of Total grants from the Ministries of Public LocalTaxes,Fees 8,956 43% Works, Local Government, and Transport, business tax 2,185 11% construction fee but not including grants from Ministries of 1,872 9% cleaning fee 1,022 5% Education, Health, or Labor) were propem (building)tax 938 5% abolished and replaced by an general grant ag. land tax 387 2% not subject to earmarking. other fees 2,55 1 12% Unconditional transfers 9,893 48% 22. The total amount to be allocated Shared Taxes 1,075 5% to the unconditional grant is determined vehicle 581 3% annually in the course of the budget property sales 495 2% process. The amount is largely based on Non-tax Itevenue 725 4% the previous year's allocation, adjusted on revenues 20,650 100% the margin to reflect the central government's overall budget constraints and the strength of competing claims. Since the system was first introduced, the formula used to distribute the grant has been changed frequently. It has also become increasingly convoluted. The current system for distributing the grant is described in Box 1. 23. There has been considerable discussion over whether local governments have enough money. Critics have noted that the level of spending by local governments in Albania is considerably lower than in other European countries-both as a percent of GDP and in relation to central government spending- and argue for an increase. As noted earlier, local governments' revenues and their share in total spending increased significantly, particularly in the first part of this decade. Given their current functional responsibilities, however, further increases do not appear to be justified. 24. There is no precise way to calculate the 'right' level of funding for local governments. Local government advocates argue that public resources should be decentralized. They note that local officials are better acquainted with local conditions and priorities than their counterparts at the central level. They say that local officials are more accessible to their constituents and can be held accountable for local results through local elections. But opponents of this view note that central governments need to retain the lion's share of revenues in order to c a w out their own functional responsibilities. This is a particularly telling argument in Albania, where the central government is responsible for financing most of the costs of education and health care. Both sides might be satisfied with an increase in the overall level of public spending. Taxpayers, however, are likely to resist this view and rightly so. See Burki, Perry and Dillinger. 1999.Beyond the Center: Decentralizing the State 6 percent of GDP (Georgia) to 33 percent (Denmark.) For Albania, the figure is 4.3 percent. But these figures are not strictly comparable. As shown in the breakdown of expenditure composition, a large part of local spending in the comparator countries is devoted to education. In most of the countries at the right end of the chart (except Italy) teachers' salaries are paid through local budgets. In Albania, they are not. By the same token, health care spending constitutes a large proportion of local spending in the more decentralized countries--particularly in Italy and in the Scandinavian countries--but is not a local functional responsibility in Albania. Very high levels of spending on social protection also inflate the level of local expenditures in the Scandinavian countries. 27. If the costs of education, health care, and social protection are excluded from the calculation, Albania is not so great an outlier. Local spending on non-social sectors in Albania, at 2.7 percent of GDP, is not so wide of the European comparator average (4.6 percent of GDP). It is quite close to the levels of local spending in these sectors as Lithuania, Bulgaria, Estonia, and even Russia. It should be emphasized that like all cross country comparisons, this does not suggest that the level of local funding in Albania is ideal. It merely shows that Albanian practice is consistent with the practices of neighboring countries. This nevertheless weakens the case for increasing the overall level of resources. Figure 3: Local Expenditure as % GDP European Comparators 28. Debates are also under way over the specific tax instruments assigned to local governments-particularly over the potential yield of the property tax and the merits of assigning local governments a share of the personal income tax. Small Business Tax 29. Albania is unusual in that the principal source of local tax revenue is the small business tax. This imposes a fairly serious administrative burden-small firms are difficult to tax-and leave the tax base vulnerable to economic success: as the economy grows, small businesses will 'graduate' into the central government's VAT base. 30. The small business tax is imposed in a fairly rudimentary way. It applies to legally registered businesses whose turnover falls beneath the threshold for the VAT. The tax rate schedule classifies firms into different sectors (about 20) and different levels of turnover (LEK 0-1000, LEK 1001-2000, etc.) and then imposes a percentage rate on the midpoint of each category, plus a fixed amount which varies according to sector. In principle, revenues could be increased by increasing the rate of the business tax. This, however, runs the risk of driving small firms into the informal sector. While efforts should continue to be made to improve the coverage and collection rate of the tax, the rate itself should not be increased. Property Tax 31. In principle, local government revenues could be increased by making more use of the property tax. The property tax (or more precisely, the building tax) is now administered by local governments. It yields very little money. Revenues in 2007 were equivalent to about US$3 per capita, or about 0.1 percent of GDP. This is in part because it suffers from the administrative weaknesses common in transitional economies. Coverage is low--properties are listed on the rolls only if their owners declare them or if they are discovered in the course of title registration.' Tax liabilities are based on property characteristics reported by owners, which are presumably also under-reported. Collection efficiency is low. Receipts are said to equal only fifty percent of the amount billed. 32. But the most striking constraint on the building tax is its rate. Parliament sets the rates on the building tax. These are based on property characteristics: a fixed amount of leks per square meter, which varies according to location and year of constru~tion.~At present, these rates are symbolic. The rate on buildings built after 1993 in Tirana, for example, is 30 leks (3 US cents) per square meter. The rate on pre-1993 buildings outside of major cities is five leks (112 US cent) per square meter. Thus an older one- hundred square meter house in a small town faces an annual tax bill equivalent to US$ 5. The Government has proposed legislation to roughly double the rate of the building tax and to introduce a 'urban plot' tax at about half the level of the building tax. Even after such an increase, individual tax liabilities would remain extremely low. 33. A precondition for any significant increase in the property tax rate is an improvement in the fairness with which the burden of the tax is distributed. At present, the burden falls disproportionately on those few taxpayers who declare their property, report its characteristics accurately, and pay what they owe. Improvements in fairness will require improvements in coverage, valuation, and collection efficiency. To this end, the Government has hired a consultant to advice on procedural reforms in the building tax. Under his terms of reference, the consultant is to focus primarily on improving coverage and valuation. With respect to coverage, the aim is to shift from the current system--based largely on owner-declarations--to one based on maps. This will permit all taxable properties to be accounted for. With respect to valuation, the aim is to shift to a market-based approach. In principle, this will allow a more precise calculation of the relative value of different properties. 34. Market-based valuation will be somewhat difficult in the current environment, as prices paid in real estate transactions are often understated in official documents. The consultant's earlier report on property taxation has concluded, nevertheless, that sufficient information on actual transactions 'A senior official of the Tirana municipalityestimates that only 58 percent of taxable properties are on the rolls. 8The current rate schedule differentiates between three geographical categories (TiranaBurres, 12 other cities, and all others) and date of construction (pre- and post 1993).Local governments have the authority to adjust the rate within individual neighborhoods within a range of 30 percent of the rate set by Parliament. exists to permit the market based approach to su~ceed.~As is standard international practice, mass appraisal techniques would be used to extrapolate from recent actual transactions to the tax base as a whole. 35. Increased coverage and market-based valuations will increase the taxable base, but may not necessarily increase actual revenue collections. The yield of a property tax is the product of four factors: (1) the number of properties listed on the tax rolls; (2) the valuations on each property; (3) the tax rate; and (4) the collection rate. Failures at the latter two stages can offset any gains made in the former two. Property taxes (particularly taxes on residential property) are acutely politically sensitive-far out of proportion to the amount of revenue they raise. It is not unknown for local governments to fail to adopt new property valuation lists for fear of political reprisals. Where local governments control the tax rate, they have been known to dramatically reduce rates in order to avoid increasing tax bills. Collection enforcement may also deteriorate, as local governments fear enforcing collections on prominent property owners. Efforts to improve coverage and shift to market based valuations may therefore pay off, but only if equal effort is paid to raising the nominal rate and improving collection enforcement. 36. Some sense of the potential yield of the property tax can be gained by examining the experience of other countries. Figure 4 below shows the yield of local property taxes, as a percentage of GDP and as a percent of total government tax revenues, for a range of European comparator countries.I0 Even in countries where the tax is well administered such as Switzerland or Denmark, it is generally not a major source of revenues. In the countries shown in the chart, local property taxes yield an average of 0.7 percent of GDP and contribute only 1.85 percent to total government tax revenues." This suggests that even if the property tax in Albania were to eventually reach the average level of the European comparators, it would generate only about LEK 6.2 billion. Although this is nearly a five fold-increase over the current yield of the property tax, it would represent only a thirty percent increase in the total level of discretionary revenues. 9Hilton, Andrew. 2006. Albania: 'Mission Report July 2006 Albanian Land Management and Urban Development Project (Property Tax Component)' ''~orpurposes of consistency, the figure for Albania includes the land tax as well as the building tax. The figure for total tax collections includes social contributions. Data is for 2006, except for Belgium, Estonia, Finland, France, Hungary, Ireland, Italy and Romania, where it is for 2005. I IFrance is a conspicuous outlier. There, property tax yields are equal to 3.8 percent of GDP and eight percent of total tax revenues. Figure 4: Property Tax Revenues European Comparators 37. Pending the completion of improvements in coverage and valuation, certain interim measures can be taken to increase property tax revenues. A pilot test of current proposals in a single pilot municipality is being planned under the LAMP project before being rolled-out to the country as a whole. The process is expected to take up to three years. This will allow the feasibility and costs of different approaches to be tested. In the interim, there are several measures that could be taken. First, Parliament should enact the proposed increase in the building tax rate, in order to raise tax liabilities to a level worth collecting. Local governments can also take action to improve collection efficiency. Sorting records to identify major delinquents, combined with conspicuous enforcement of penalties, can produce a major one-time increase in collections on arrears, as well as having a salutary effect on collection efficiency in the future. Consideration should also be given to expanding the range of rates local governments can impose. As noted earlier, local governments are permitted to impose rates 30 percent above or below the rate set by Parliament. The ceiling could be increased to, say, 60 percent. This would permit local governments that have the political will to impose significant levels of property taxes on their constituents to actually do so. Box 2: Juggling the Costs and Standards of Property Tax Administration Efforts to improve property tax administration have to take costs and standards into account. Mapping can be expensive. It can also be extremely slow if tied to the title adjudication process. The consultant now employed by the Government recommends that boundary and ownership data be extracted ftom property registration data maintained by the national property registry, IPRA. But the IPRA records are far from complete and may remain so for some time. It can take years to adjudicate property titles, particularly where there has been a history of irregular property transactions and informal development. Tying property tax administration too closely to title records could severely delay improvements in coverage.It is also unnecessary, since precise parcel boundaries are not required for property tax administration and property tax liabilities can be assigned to the property rather than to the owner. The consultant's prior report on property taxation proposes an alternative, more rough-and-ready approach to assembling a comprehensive list of taxable properties, based on street addresses. This proposal should be given serious consideration. It could also be tied into the street addressing program of the World Bank-financedLAMP project. Personal Income Tax 38. An alternative to the small business tax or increased reliance on the property tax would be to share a proportion of the personal income tax (PIT) with local governments. In principle, this would offer local governments access to a broad based, more easily administered, and less politically sensitive method of taxing their constituents. As shown in the chart below, many--although by no means all--of the European comparator countries assign local governments a share of a nationally administered personal income tax. In former Socialist countries, this is generally distributed on an origin basis. The percentage of the PIT that is shared with municipalities varies among them, ranging from 30 percent in the Czech Republic and Poland to 73 percent in Latvia and 94 percent in Slovakia. Figure 5: Local Personal Income Tax Revenues European Comparators 39. The Government's current decentralization strategy, proposes allocating up to 15 percent of the revenues from the personal income tax to local governments.'2 According to the timetable provided at the end of the report, this reform was to be in place by January 2007. It is not yet in place and the impetus behind it has apparently dissipated. This isjust as well. 40. Sharing the PIT would not be a particularly equitable way of financing local government. The vast majorities of PIT revenues in Albania is derived from payroll deductions, and are thus generated only in jurisdictions that have large numbers of people employed in large, formally-registered organizations. Such organizations tend to be concentrated in Tirana and a few other large municipalities. In 2005 (the last year for which complete data is available) the Tirana district alone accounted for 63 percent of all PIT revenues. Full half of the 36 tax administration districts yielded less than LEK 20 million per district. Allowing local governments to retain some or all of the PIT generated within their boundaries might provide Tirana, Durres, and a few other large jurisdictions with an effective means of self-taxation. But outside of the big cities, the burden of the PIT would presumably be confined to public 12~hisrepresents a fairly modest increase in the overall level of local revenues. Based on revenue data for 2007, 15 percent of the PIT would be equal to about 2.22 billion leks, or about ten percent of local discretionaryrevenues. employees--e.g. school teachers--and the occasional employee of a branch office of a firm headquartered in Tirana. 41. In the long term, as the base for the personal income tax expands in more jurisdictions, there will be a case for personal income tax sharing. This should be implemented in the context of broader reform in the local tax structure, however, including the possible elimination of the local business tax and after the efforts to improve property tax yields have been tested. For the present, the small business tax and the property tax, along with construction fees and unconditional transfers, provide a sufficient number of instruments to finance the expenditure responsibilities of local governments. FINE-TUNING TRANSFERSYSTEM THE 42. Albania's system of unconditional grants has many desirable properties. The principal criteria used to distribute the grant are transparent, readily verifiable, and reasonably equitable. There is little discretion (other than in the designation of 'distressed municipalities') in the measurement of these criteria. But for all the reforms that have taken place since 2000, the current system of local government transfers still has some weaknesses. First, it is unstable. And some would argue that it is insufficiently equalizing or insufficiently earmarked for critical expenditures. Stability 43. The key weakness in the transfer system is its unpredictability. This is partly because the overall level of the transfer is subject to annual budget negotiations. It is also because the formula for distributing the transfer changes from year to year. The result is a high level of uncertainty, which undermines local budgeting, and a potentially excessive role for short term political considerations in determining the distribution of funds. 44. The overall level of unconditional grants has changed fairly radically since it was introduced (See Figure 6). In nominal terms, unconditional transfers rose from LEK 0 to nearly LEK 10 between 2000 2002, to Figure 6: Trends in Unconditional Grants by 30 percent in the following year. They have subsequently returned to their former 12000 level (in nominal terms). There have also been changes in the distribution formula. IOOOO The share allocated to regions (as a group) 8000 has fluctuated from nine percent to nineteen percent, with corresponding changes in the Y 6000 local governments' share. (It has, however, 9 remained at nine percent over the last four 4000 years: 2005-2008.) While the importance 2000 of population in distributing the local government share has remained relatively 0 ,. ,. ,, constant--ranging from 63 percent to 70 c n8 E8 gZ g8 g 8g 8g 8o 8o 6 8 o W W N percent since 2002--the division of the remainder between communes and municipalities has shifted. In 2002, the communes' share (of the local government share that was not distributed on the basis of population) was only ten percent; with sixty percent assigned to municipalities and an additional thirty percent to Tirana. As of 2007, the communes' share has increased to 50 percent, with the municipalities (including Tirana) receiving the other half. (The special share for Tirana was eliminated in 2005.) 45. Some degree of flexibility in the level of transfers is certainly desirable. Subnational governments, as a group, have to compete against other central government priorities for funding. The transfer system also has to be able to adjust to changes in the overall tax structure. It is no coincidence that the major drop in unconditional grants in 2003 occurred at the same time the Government increased the level of local taxes. Future functional decentralization will also require a reconsideration of the overall level of unconditional transfers. 46. Nevertheless, the Government should consider enshrining the rules governing unconditional transfers in a more permanent form of legislation, such as the local government organic law or the organic law on budget preparation. This has been the practice in other Eastern European countries. Poland and Serbia, for example, both fix the parameters of their respective transfer systems in a separate local government finance law. Such legislation should establish a mechanism for determining the overall amount of the transfer each year-as a fixed share of aggregate national tax revenue, for example. The legislation should also specify the division of transfers between classes of jurisdictions (regions, municipalities, and communes) and the criteria to be used to distribute funds to each jurisdiction within each class. To provide for flexibility, these provisions could be subject to periodic review at, say, five year intervals, by a high level group including representatives of the relevant ministries at the central level and the associations of local governments. Equalization 47. The present system of tax assignment and transfers yields a high level of variation in per capita revenues among individual jurisdictions. In 2007, per capita revenues from all discretionary sources ranged from LEK 2,432 (in Kastrat) to LEK 14,458 (in ~eskovik).'~The dimension of disparities is illustrated in the chart below, which shows the per capita revenues of each jurisdiction as a single vertical line, with local governments ranked from left to right from poorest to richest. As shown, there is a relatively small group of local governments with per capita revenues of over LEK 8,000. These include Tirana, Saranda and several small jurisdictions. (Other major cities--Durres, Fier, Shkoder, and Elbasan-- have per capita revenues in the range of LEK 5700-7500.) The majority of local governments have discretionary revenues ranging from LEK 2000 to LEK 7000 per capita. 48. This variation is largely due to variations in the per capita yields of local taxes and fees. The vast majority of revenues from taxes and fees is generated (and kept) in the large cities: particularly Tirana and Durres. Tirana derives 84 percent of its discretionary revenues from these sources; Durres 78 percent. At the same time, many smaller jurisdictions derive virtually nothing from local taxes and fees. 239 of the 374 jurisdictions generate less than LEK 1,000 (US$ 10) per capita from these sources. Overall, the standard deviation in per capita local sources is LEK 3,307, nearly 2.45 times the average. I3~hisanalysisexcludesthree extreme outliers:Kashar, Dajt and Farke. Figure 7: Variation in Per Capita Revenues I All LGs in Rank Order, 2007 49. This variation is offset, to a large degree by the unconditional grant, including the fraternal14 equalization of the small business tax. Because the grant is distributed largely on a population basis and because the amount of the grant is considerably larger than the revenues generated from shared taxes, local taxes, and fees, it has a dramatic effect on inter-jurisdictional disparities. As shown in Table 3, below, the coefficient of variation in per capita discretionary revenues drops to -70 once unconditional grants are taken into consideration. 50. Although some would argue that inter-jurisdictional Table 3: Indicators of Variations in Per Capita Discretionary disparities remain too wide, Revenues there is not a strong case for Local Total Taxes Unconditional Discretionary changing the transfer formula to and Fees* Grants Revenues increase equalization. The case Standarddeviation 3307 1429 3383 for equalization would be fairly Average 1346 3490 4836 compelling if local governments Coefficientof variation 2.46 0.41 0.70 relied on discretionary revenues to *includingshared taxes finance functions with important distributional implications, such as teachers' salaries, the ndihma ekonomike or the operating costs of primary health clinics. One could argue that all Albanians have the right to a basic level of education, health care, and social assistance regardless of the tax base of the jurisdiction in which they live. But in Albania, such expenditures are financed by the central government, either in the form of conditional transfers (such as the former transfer for teachers' salaries and the ndihma ekonomike) or through direct payments to facilities (as is the case for primary health). An increase in the redistributive element of the transfer formula would instead redistribute funding that is largely used for roads, solid waste management, street cleaning, and various undefined administrative services. In the absence of any major 14I.e.,the arrangement by which local governmentswith high levels of business tax revenue per capita are required to share some of it with poorerjurisdictions. redistributional functions to be financed from discretionary revenues, the case for leaving the current level of equalization as it is compelling. Earmarking 5 1. There is, finally, some ongoing debate over the merits of re-introducing earmarked funding. For the present, this debate concerns only two sectors: roads and education. As water companies are transferred to their respective municipalities, it could also include the water supply sector. 52. Funding for the maintenance of regional and local roads has been erratic. Central government direct support has varied from year to year. Local governments tend to devote much of their discretionary resources to other sectors or to general administration. The unreliability of funding for maintenance affects not only the condition of the network but the level of capital investment in roads as well. As described in the World Bank report, Policy Optionsfor the Local Roads Sector (July 2007) capital investment in the national and local road sectors now comes primarily from donors. Donors are increasingly reluctant to provide financing for capital investment unless a domestic system for financing maintenance is in place. 53. The recent ANTP reportI5 recommended the creation of a separate roads fund at the national level, into which road-user-related taxes would be placed and used to finance expenditures in the sector. In principle, this would also be an option for the local road sector. The recent Policy ~ o t eanalyzed this ' ~ option, but recommended the current budget-based approach. Earmarking, clearly, reduces the flexibility of democratically elected bodies to choose their spending priorities. In the absence of earmarking, local governments are able to spend their discretionary resources as they choose. Their choices may be technically unattractive-local democracy in Albania does not necessarily focus resources on activities with high economic rates of return. But re-introducing earmarking for road maintenance flies in the face of recent trends toward increased local autonomy. It should be seen as a last resort. 54. In any event, there is some prospect that the problem can be partially solved through other means-particularly the proposal to transfer lifeline rural access roads to the central government and abandon what are essentially private access roads for forestry and mining enterprises." This should free local government to spend more on the lightly traveled feeder roads that would remain under their jurisdiction. 5 5 . The case for earmarking funding for education rests on the belief that local governments willfully spend too little in fulfilling their obligations in the education sector. As noted earlier (and in the separate policy note on education financing) local governments are responsible for financing the non- salary costs of school operations (such as heating and lighting) as well as minor repairs, such as the replacement of broken windows. Major construction, such as the building of new classrooms, is financed out of separate capital grants. While there is as yet no evidence that heating and lighting are "~ouisBerger (2005), Volume 3, pp. 3-24. 16World Bank. July 2007. Institutions, Management, and Financing of the Local RoadSector: Optionsfor Albania I7~orrural roads, there are three informal classes of road: (i) local access roads - the key 'lifeline' regional access roads, which carry more traffic, (ii) private access roads which are essentially private roads for forestry and mining enterprises,and (iii) other local, or commune, roads which connect lifeline roads to outlying communities and areas. Following from recommendations of ANTP, the first would be transferred to the GRD, whilst the upkeep of the second would be effectively left to their owners. Only the third would remain the responsibility of the relevant local government unit. systematically under-funded, it has been noted that many schools are severely deteriorated, with collapsing roofs and gaping holes in their floors. 56. Whether earmarking is the solution to this problem is not obvious. Unlike heating and lighting costs, school rehabilitation is not a recurrent cost. Nor is physical deterioration uniformly distributed. Schools were built at different times and have been subject to different degrees of neglect over the years. The urgency of major school repairs also varies according to trends in enrollment. (One would not want to spend large sums rehabilitating schools that are likely to be closed.) Under these conditions, earmarking may not be appropriate. Instead of compelling all local governments to spend a fixed share of their revenues on school rehabilitation, it would make more sense to finance such expenditures through capital grants subject to case by case evaluation. 57. Further plans for functional decentralization could have important fiscal implications for local governments. These include the ongoing transfer of water companies to local governments and longer term proposals to increase local discretion over the allocation of education spending. These may require adjustments in the overall level of transfers to local governments andlor the redesign of certain transfers to ensure that central government objectives are effectively carried out. Water Company Decentralization 58. The transfer of ownership and management of Albania's water companies are now well under way. Under the municipal organic law, water supply is a local function. But until recently, urban water supply was provided by 54 companies, all but nine of which fell under the supervision of the central government's General Directorate of Water and Sanitation (GDWS, a department within the Ministry of Public Works, Transport, and Telecommunications) and were legally owned by the Ministry of con om^.'^ The Government is in the process of transferring ownership and supervisory responsibility for these companies to the local governments they now serve. Since the majority of companies serve more than one jurisdiction, shares of a given company will be transferred to the jurisdictions it serves in proportion to their respective populations. Thosejurisdictions will in turn form a shareholders association to provide overall oversight of each company. The shareholders association will appoint a standing supervisory board, which will, inter alia, appoint the managing director of the company. 59. However, the process, which began in December 2007, is proceeding somewhat slowly. As of March, 2008, the tasks of constituting a shareholders assembly and nominating the members of the supervisory body had been completed only in eleven water companies. In twelve, an assembly had been constituted but the members of the supervisory body had not yet been nominated. In the remainder, the assembly of shareholders had not yet met. These include the water companies serving Tirana and Durres. Tirana, in particular, has declined to accept ownership of its water company until a controversy over the company's boundaries has been resolved. 60. The transfer of ownership in the water companies could have major financial implications for local governments. At present, the water companies are dependent upon central government subsidies to cover the costs of current operations. These take two forms. The first is a quarterly subsidy allocated by the GDWS. As described in Box 3, the amount of the subsidy is based on the difference between the operating costs and the operating revenues of each company, with operating revenue based 18Some communes had isolated systems in villages, which they operated on their own. 16 on billings rather than actual collections. This amount is adjusted on the basis of quarterly monitoring indicators to determine the actual amount of the quarterly subsidy. The adjustment process is far from transparent, however. The second operating subsidy consists of end-of-year direct payments to the national power company, KESH, to cover the water companies' unpaid electric bills. 61. The current system for allocating subsidies provides no incentive for water companies to reduce unaccounted-for water, economize on staffing or other operating costs, or increase tariffs. As a result, the scale of operating deficits could rise in the future. In the short term, the Government intends to continue to finance the water companies' operating deficits. But it is committed to phasing out its own support over time. If the water companies are unable to reduce operating deficits as fast as the Government reduces operating subsidies, then the task of providing subsidies will fall to the respective local governments. As shown in Table 4, total operating subsidies in 2007 were equal to Lek 1.646 billion or about 45 percent of the direct operating costs of water and sewerage services. Although this is not particularly large as a proportion of central government expenditures (0.62 percent of the total, including social security), it is equivalent to eight percent of local government discretionary resources. Box 3: Calculating the Operating Subsidy for Water Companies At present, the subsidy to each utility is based on its 'calculated deficit'. This is determined, for each utility, by: (1) calculatingthe average cost per cubic meter of water sold (based on actual expenditures);(2) subtractingthe average tariff per cubic meter; and (3) multiplying the result by the number of cubic meters sold. The amount of downward adjustment to determine the actual amount of the quarterly subsidy is a matter of judgment but reflects: (1) ratio of the subsidy request to the calculated deficit (presumably smaller requests are favored); (2) the size of the water utility (smaller utilities are favored); and (3) whether the system is gravity based or relies on pumping (the latter are favored). 62. The current proposal to reform the system of operating subsidy is not yet well-articulated. Under a Table 4: Current Operating Subsidies to proposal currently under consideration, the two Water Companies (LEK mn) subsidies-the subsidy administered by the GDWS and the Total To water To companies KESH** direct transfer to KESH--would be combined into a single 2005 1,673 780 893 subsidy, administered by the GDWS. The overall level of Government support would decline by ten percent annually 2006 1,483 780 703 (based on its 2005 base level). While there is consensus on 2007 1,312 790 522 the aggregate rate of decline in subsidies, the Government is still considering alternative mechanisms for allocating *Budgetrequest. these cuts among individual companies, and in the process, ** Transferto KESH to cover water providing incentives for improvements in financial results. companies' current power arrears This issue should be addressed as a matter of urgency. 63. Under the most recent proposal, the GDWS would continue the current gap filling approach, but would add a 3-element performance indicator, reflecting: (1) the percentage of non-revenue water; (2) the collection rate (tariff revenuelbillings) and (3) the percent of costs recovered through tariffs. A company's performance against these criteria would be used in assessing the proportion of its deficit to be covered by operating subsidies, but would not affect the level of such subsidies in any direct, mechanical way. 64. An attractive alternative would be to simply reduce the operating subsidy of each company by a uniform amount. Here is why: to provide an incentive to reduce operating deficits, any allocation mechanism must begin by imposing a clear constraint on the overall level of subsidy to each company, rather than continuing the gap-filling approach. In theory, it might be desirable to establish different targets for subsidy reductions for each company. Those with high levels of non-revenue water or low tariff collection rates, for example, could be given higher targets. But this would require a high degree of discretion and authority on the part of the GDWS. This discretion may be difficult to employ impartially. In order to avoid charges of discrimination or favoritism, it is best that the target be uniform for all companies. 65. It is therefore recommended that the operating subsidy be allocated on the basis of the amount granted to each company in a base year (such as 2005) reduced by a fixed percentage. As the MOF is committed to a ten percent annual reduction in the overall level of subsidies, this percentage would presumably be ten percent. Government payments to KESH on behalf of the water companies would be included in this calculation. 66. This approach has several advantages. It avoids possible charges of discrimination or favoritism among the companies. It also ensures that the overall level of subsidies is consistent with the aggregate subsidy level agreed with the MOF. It is true that in taking the 2007 level of subsidies as a base line, it rewards companies that have a history of poor management. One can only take comfort in the fact that as the overall level of subsidy declines, the arbitrariness of this approach will decline as well. As the level of operating subsidies drops to zero, all companies will be forced to cover their operating costs solely from tariff revenue. Managing Teachers' Salaries 67. Over the longer term, the Government is considering changes in the way primary and secondary education are managed and financed. These changes could include a shift to an enrollment- based formula for distributing funds for teachers' salaries. In theory, this could improve equity and provide an incentive for a more efficient configuration of the school n e t ~ o r k . 'At ~ present, schools in rural areas tend to be under enrolled while schools in Tirana and other growing cities are over crowded. Enrollment-based financing could reduce disparities in funding per student and force under enrolled schools to close or consolidate. As described in the Education Policy Note, however, an immediate shift to enrollment-based financing for teacher's salaries could result in teachers going unpaid in some jurisdictions. If the benefits of enrollment-based financing are to be realized, the authority to match the number of authorized teaching positions to enrollment will first have to be transferred, either to the regional offices of the Ministry of Education, the local governments, or to school directors. As discussed in the education policy note, such a reform should be undertaken in conjunction with broader reforms in the education sector, if it is undertaken at all. 68. The main weakness in the Albanian structure of local finance is not the system for financing recurrent expenditures but rather the system for allocating funding for capital works. The Government is experimenting with a variety of approaches to financing local capital works. While these show much promise, further reforms are required. Competitive Grants 69. Municipal capital investments are financed from a variety of sources. Donors finance municipal projects, either directly in individual jurisdictions, or through intermediaries such as the l9See World Bank. "Per Student Financing of General Education in ECA: Has it Delivered on its Promise?" (forthcoming) Albania Development Fund (see below). The Government provides counterpart funding for these projects and funds other projects directly from the budget. Since 2006, the Government has been experimenting with a system of competitive grants to finance small scale local investments. While this shows considerable promise, several issues remain to be addressed. 70. Until 2006, central government funding for small scale local investments-in schools, health facilities, water supply, and roads--were allocated by sectoral ministries in the form of conditional grants. Project selection procedures were controversial. According to the Government's Economic and Fiscal Program, 'the analysis of the distribution of investments among different regions, communes and municipalities for the years 2000-05, reveals a systematic absence of objective criteria for distribution, unpredictability and lack of transparency.' 71. In 2006, the Government introduced a new scheme. Investment funds for education and primary health were allocated to the regional governments, with the understanding that they would be distributed among their respective local governments on the basis of regional priorities. This approach is now considered a failure. Rather than concentrating investments on regional priorities, the regions distributed small amounts to almost all local governments. 72. Also in 2006, the former system of conditional grants for roads was abolished and replaced by a system of competitive grants. Annex 4 of the 2006 budget law set out the terms of the competitive grants: 'State budget funds allocated to line ministries that finance projects in fields (that) are local government own functions, shall be allocated, monitored and controlled in cooperation with local government representatives. Ministries are to allocate funds to local governments based on an open and transparent competitive process, according to criteria set forth in this annex.' Grants in this category are no longer confined to roads, but can be used for any local function, as defined by Law no. 8652. 73. The Annex provided general criteria for allocating the competitive grants. These are: The rate of expected impact on economic and social development, as well as the rate of compliance with local and lor regional priorities of development; The rate of impact on poverty reduction and on the increase of access over basic services; The extent to which the project promotes cooperation between LGUs; The rate of community participation in decision making and counterpart funding; The extent to which the grant would provide counterpart funding for foreign-funded projects; The technical qualities of the project. Preference was also given to ongoing projects, for which there are contract obligations. 74. In 2007, the competitive grant scheme was expanded to include investments in the sectors previously assigned to the regional governments--education and health--and to some water sector investments. The general criteria for funding allocation remained unchanged. The procedures for project selection, however, were altered. Rather than relying on specific criteria, allocations were to be made by 'inter-ministerial committees for evaluation' consisting of representatives of the relevant ministry, the MOI, the MOF, and the local government associations.' As of 2007, the competitive grants system therefore governs local capital investments in four sectors: education, health, water supply and 'general municipal infrastructure'. As shown in Table 5, the Government budgeted LEK 1.6 for education, LEK 800 million for health, LEK 1 billion for water and LEK 1.9 billion for general municipal infrastructure in 2007. 75. Each of the four competitive grants Table 5: Budget Allocations for Competitive Grants employs a slightly different approach to (LEK mn) project selection. 2007 2008 Education 1,600 2,029 General Municipal Infastructure Health 800 0 Water Supply 1,000 1,000 76. The procedure for allocating General Municipal Infrastructure 1,900 1,664 competitive grants in the general Other 0 100 municipal infrastructure category begins ~~~~l 7,307 6,801 with a call for project proposals from all local governments. These proposals are referred to the Albanian Development Fund (ADF), which acts as technical secretariat to the Ministry of Interior, the ministry nominally in charge of this sector. ADF performs a technical evaluation of the proposals. As discussed in Box 4, ADF's evaluation system is essentially an elaboration of the general criteria set out an annex to the annual budget law. Once evaluated, proposals are referred to the inter-ministerial evaluation committee, consisting of representatives of the MOI, the MOF, the association of municipalities and the association of communes. The committee selects a final list of projects for financing. 77. In 2007, the program got off to a slow start. The total amount of financing for the competitive grant was not determined until well into the budget year. Nevertheless, within the available time span, a total of 750 project proposals were received. Of these, roughly 400 were immediately rejected because they lacked detailed designs. Of the remainder, another 150 were rejected on the grounds that the jurisdiction had already submitted at least two proposals. The remainder survived the evaluation process and was forwarded to the inter-ministerial committee. 78. ADF's evaluation work appears to have been carried out in a transparent and professional manner. It is notable that it began with a mechanical division of resources among the regions (with population a major factor) and that the criteria specified in the budget law were further elaborated into a detailed set of questions and measures. ADF evaluations, in turn, appear to have largely determined the choice of projects that were ultimately approved for funding. The total package of approved projects was sufficient to exhaust the available funds. The inter-ministerial committee neither deleted projects or added any of its own. 79. One immediate issue, nevertheless, needs to be addressed. At present, ADF and the inter- ministerial committee approve too many projects. As a result, even though the projects are small, they are not fully funded. Construction proceeds haphazardly and is subject to costly interruptions. According to the ADF, the backlog of approved but uncompleted projects is so large that it would require two to four years to complete at current levels of funding. For 2008, ADF recommends that the majority of funds be used to complete projects approved in 2007. This is an admirable, albeit short-term, suggestion. Over the long term, it would make sense to tighten the criteria, in order to reduce the number of projects a level commensurate with the amount of available funding. 80. Part of ADF's success appears to derive from its long-standing role as an intermediary for donor funds. Staffs are well paid, well trained, and familiar with donor-supported capital allocation mechanisms. This is not, of course, a permanent solution. Whether the process would survive the transfer of ADF's functions into the normal structure of government--with its reduced salaries and increased proximity to the political process--is not clear. Should this occur efforts should be made to retain ADF's technical strengths. Albania could, for example, employ consultants to evaluate projects. Some countries employ specialized staff in the Ministry of Finance. The transparency of the process could be maintained by posting the results of the selection process on a government website. An administrative appeals mechanism could be introduced, which would allow local government submitting proposals to appeal the evaluation committee's decisions on procedural grounds. 81. The Government should impose a more sophisticated project appraisal approach to larger projects. There is an extensive literature on project evaluation. Much of this is justly criticized for advocating approaches that are unrealistically complicated. In theory, it might be desirable for every project to emerge from a comprehensive and participatory analysis of local development needs, carefully integrated with the plans of national ministries and other stakeholders and then appraised according to rigorous economic, social, and environmental criteria. But this is likely to be administratively impractical. Albania could, however, develop a 'cookbook' for project evaluation, with a specific evaluation methodology for individual sectors. The EU offers one such example in its guidelines for accessing structural funds. Any such methodology should be coordinated with the Department of Investment Management's (DPIM, a department in the Ministry of Finance) new public investment management procedures, which applies to investment submissions starting in 2008. Box 4: ADF Criteria for Evaluating Municipal Infrastructure Projects The 2007 Budget Annex specifies that the Albanian Development Fund (ADF) is to 'assist the committees through preliminary assessments and rankings of projects'. ADF's approach began with a deliberate outreach program, starting with a campaign in each village of each commune and each neighborhood in each municipality. The councils of each commune and municipality were instructed to select their top three priorities. Funding for municipal infrastructure was then divided on a regional basis according to population and poverty rates and a scoring system was used to rank the projects in each region. ADF's scoring system assigns a range of points to each of the eight general criteria specified in the budget annex, with a maximum score of 140points. It also provides a list of 22 detailed questions to be addressed in determining the number of points to be assigned under each criterion. For example: Rate of expected impact in the socio-economic development (40 points) Does the road have an impact specifically on tourism development, industrial or agricultural production and trade? A road directly connected or situated on an industrial construction zone, scores higher than an internal road. Impact level onpoverty reduction and access growth at the basic services (25points) what percentage of households receives the ndihma ekonomike from the Government in the relevant village/neighborhood? Number of direct or indirect beneficiaries (25points) what is the population of the beneficiary village/neighborhood. Points are added for villages with growing populations. Cross border cooperation: 20points Is the projectjointly proposed by more than one local government unit? Participation level of the community in decision making (IOpoints) does the financing plan meets requirements for counterpart contributions? Do local council records define the project as a priority? Are there documents demonstrating that local communities have participated in the decision making on the priorities? Contribution to counterpart funding (IOpoints) would the project help provide counterpart hnding for a donor financed project? Continuity (IOpoints) is the project already under implementation, with ongoing contractual obligations? The scoring page is attached to every reviewed project, along with an explanation for each question, photos of the on- site current situation, and a technical report. Only small projects are subject to scoring. Requests for finding to continue projects proposed by municipalities (as distinct from communes) are not scored. Water Supply 82. The competitive grant for water supply has, until recently, been largely focused on small investments in the few water supply systems owned by municipalities and communes. Large scale investments, and investments in companies owned and managed by the central government, have been financed directly by donors and/or through project-specific support from the central government budget. With the decentralization of the water companies to local governments, the importance of competitive grants is likely to increase. Plans are also afoot to introduce performance incentives into the grant allocation mechanism. 83. At present, requests for funding under the competitive grant are submitted to the GDWS, which evaluates them, first, on technical grounds.20This results in a dramatic reduction in the number of candidate projects. In 2007, roughly 1,000 proposals were submitted. Of these only 110 passed technical muster. These were then subjected to the Annex 3 criteria. Only 70 survived this screen. These were sent on to the inter-ministerial evaluation committee, which approved them For 2008, the inter- ministerial committee requested a larger batch of project candidates so they could have more leeway over the choice of projects. 84. The Government is now considering changes in the criteria used to allocate the competitive grant for water. These changes are intended to increase transparency, while introducing performance incentives into the allocation criteria and assuring a minimum level of funding for water companies that have unusually low levels of coverage or highly intermittent supply. In principle, the new formula might apply not only to the competitive grants but to larger scale Government investments in the sector. 85. A consultant has been retained to prepare a specific proposal. Under his most recent submission, the annual competitive grant would be divided into two pools: one to be allocated based on performance; the other based on need. The performance pool would be allocated in three steps. First, each company's performance would be measured on the basis of eight quantitative indictors, as shown in the Table 6, below. (Performance would be calculated as the percentage change in each period over the previous period.) The eight indicators would then be collapsed into a single performance indicator, using the weights shown in the right-hand column of the table. The composite indicator would be multiplied by the number of customers served by the particular company, yielding a population-weighted composite indicator for each company. The resulting population-weighted composite indicators would be summed for all water companies. Each company's share of the total would determine its share of the competitive grant.22The same sequence of steps would be used to calculate each company's share of the 'needs' pool, based on the indicators shown in Table 7. 20~nmuch the same way that the ADF evaluates proposals submitted in the general municipal infrastructure category. 21 This may be explained by the fact that the costs of these 70 surviving projects were barely enough to consume all the available funding. 2 2 ~ hconsultant report discusses four different methods for calculating the amount of the grant. The preferred e method (option 4) would exclude certain indicators from the calculation of a company's aggregate performance measure under two circumstances: (1) where the company's improvement against that indicator is below the national average or (2) where the company has already achieved a target performance level for that indicator. Table 6: Performance Indicators (measured as percentage change over previous period) Indicator Measurement Weight Unaccounted-forwater Volume soldvolume produced 15% Metering coverage No. of metered connections/#of total connections 10% Water coverage Population connectedtotal population 5% Wastewater coverage Population connectedtotal population Revenue collectionrate Amount collectedamount billed Overstaffing indicator Reciprocal of labor costsloperatingcost Operatingratio Amount billedoperating cost* Energy efficiency** Energy cost as % operating cost* 20% * excluding depreciation ** proposal distinguishesdifferent indicatorsfor gravity-based,pumping based, and mixed systems Table 7: Need Indicators (measured as absolute level) Indicator Measurement Weight Water coverage Difference between national average coverage and coverage in company x 20% Wastewater coverage Difference between national average coverage and coverage in company x 20% Affordability Avg, operating revenues per capitaIGRP per capita 20% Water Supply Hours Difference between national target for hours of supply and hours of supply of Supply in company x 40% 86. The result, according to the consultant's simulations, would be a fairly high degree of variation in grant allocations, as measured in per capita terms. The coefficient of variation in grants per capita (based on the performance criteria alone) would be -57. Large companies would nevertheless receive a large proportion of the funds. Tirana, with 34 percent of Albania's water customers, would receive 23 percent of the performance part of the grant. Durres, with eight percent of the country's water customers would receive 11 percent. 87. While the proposal is broadly consistent with the Government's objective, it does, however, embody certain contradictions that are inherent in any system of grants that rewards improvements in performance. It yields a transparent, objective formula for allocating capital grants, and one that favors companies whose performance is either improving or whose initial conditions are particularly poor. While some performance improvements can be made at little cost (in financial, if not political terms) others cannot. Improvements in the coverage of metering or water supply and wastewater coverage, for example, are expensive. So are reductions in technical losses (which affect the operating ratio). Under the current proposal, performance grants would be awarded to companies whose performance has already improved. As a result, there is some risk that only companies that do not need the money would be eligible to receive it. 88. There are ways around this dilemma. One would be to include only factors that are cheap to implement. This might include improvements in the collection rate (through stricter enforcement on delinquent customers) or the staffing ratio (through dismissals). In principle, tariff increases would also be cheap to implement. (Taken in isolation, however, they may not be desirable if they merely generate more revenue to cover overstaffing and other inefficient practices.) Limiting the performance criteria to low- cost reforms would, however, leave several important performance indicators-including unaccounted for water and the operating ratio-off the list of behaviors to be rewarded. Another approach would be to allocate the grants on the basis of promised performance improvements in the future. Under this approach, financing would be provided to make such improvements possible. This approach is likely to produce extravagant and unachievable commitments, however. For these reasons, the Government would be well advised to leave the performance criteria largely as they are proposed. The Government could, however, provide funding for other performance enhancing measures-particularly metering-through a separate national program. 89. The needs criteria largely make sense. In effect, they favor companies that are particularly far behind the national average (or a national standard) in terms of coverage or reliability. At first glance, the needs criteria and the performance criteria might appear mutually contradictory. Under the performance criteria, a utility would be rewarded for having a high level of water coverage. Under the needs criteria, it would be punished. The two are consistent, however, as long as the performance criterion is measured in dynamic terms-improvement in coverage levels-while the needs criteria is measured in static terms-a below-average level of existing coverage. The affordability criteria may be difficult to measure, however, as it is unlikely that accurate figures on gross regional product per capita are available for each company. 90. The use of a formula does not preclude the need for project-by-project evaluation. Presumably, the new competitive grant allocation formula would merely set a ceiling on funding. Individual projects still have to be evaluated individually. The Government will also have to decide how broadly to apply the new formula; i.e., whether to apply it to larger scale investment financing from donors or directly from the budget. The consultant recommends that the formula not apply to donor funded projects, particularly in large cities. This may undercut the impact of the performance element of the proposed formula, however. Water companies-and their new municipal owners-may not respond to performance-conditioned grants if condition-free funding is available from other sources. Some effort should be made to apply the proposed criteria to donor-financed water projects as well. Education and Health 91. The Ministries of Education (MOE) and Health act as the technical secretariats for these two sectors, respectively, in evaluating individual project proposals before forwarding them the respective inter-ministerial evaluation committees. The MOE's experience with the competitive grants is mixed.23Like the ADF and the General Directorate for Water, the MOE is inundated with proposals. For 2008 funding, it received 433. Of these, 100 passed muster and were sent to the inter-ministerial committee. Of these, 68 were approved, with another 17 put on the tentative list for further study. 92. MOE's principal concern is the absence of any system for prioritizing projects within the education sector. In effect, the competitive grant for education relies on local government to propose projects. In general, these tend to be small rehabilitation projects, rather than new schools-in part because the documentation requirements for new construction are much more demanding. By any objective measure, priorities for school investment vary widely. Schools were built at different times and have been subject to different degrees of neglect over the years. Changes in the distribution of Albania's population also affect the urgency of investment in different areas. Recent rural-urban migration has left some facilities under-used and others overcrowded. Although the MOE has defined a set of geographical investment priorities, the MOE must rely on requests from local governments to implement them. 93. The Government should consider using national sectoral investment plans in the allocation of investment funding for education. The World Bank is already supporting the preparation of a health facilities rationalization plan, which will identify which facilities to close and which to expand. A similar exercise would be appropriate for schools. This could be accomplished either by imposing an overriding 'geographical priorities' criterion on the allocation of these grants or by simply re-centralizing investment decisions in the Ministry of Education. The World Bank financed a school mapping effort under the previous education loan but it has not been used for facilities rationalization. The new education project 2 3 ~ hMinistry of Health was not interviewed for this report. e will finance a management information system. At present this is not planned to be used for capital investment planning. It should be. Municipal Borrowing 94. In the very long term, the Government could remove itself from the business of allocating funds for local capital investment and turn this function over to the private capital market. Local governments could identify their own investment priorities and seek funding from private banks. Some move in this direction is reflected in the Government's decision to remove the prohibition on local government borrowing and substitute limited borrowing powers. 95. The law on municipal borrowing that was recently enacted by Parliament has several valuable features.24First it defines debt broadly, to include any monetary obligation or liability created by a financing agreement, the issuance of securities, or a guarantee to third parties. As such it closes several loopholes which have been a source of problems in neighboring countries. 96. It also provides an appropriate level of Ministry of Finance oversight. The law provides that local governments must inform the MOF of the issuance of any debt within ten days of doing so. The MOF's responsibility, once it receives this information, is limited to the validation of the local government's compliance with procedural requirements and debt limitations (as discussed below). Failure by the MOF to respond within 20 days of notification constitutes approval. This somewhat detached approach to regulation is desirable as it limits the scope for political interference in what should be the purview of local government: the actual choice of the project to be built. 97. The law provides for closer scrutiny in three cases: debt issued to refinance existing debt, debt issued by a local government that has defaulted within the previous five years, and debt issued in international markets. In these cases, positive approval must be issued by MOF. The higher level of scrutiny in these three cases is justified. The requirement for international borrowing is particularly important. International private capital markets are not adept at distinguishing a default by an Albanian local government from a default by the central government. Under these circumstances, the central government could be compelled to assume the obligations of a defaulting local government in order to protect its own reputation. Such authorizations should therefore be granted sparingly. Ceilings on Local Government Borrowing 98. The new borrowing law sets out several quantitative restrictions on municipal borrowing. With respect to short term debt, it limits the stock of debt at any one time to ten percent of revenues from local taxes, fees and shared taxes in the previous year. By excluding unconditional transfers from this calculation, the ceiling is quite restrictive and rightly so. 99. Article 18 of the law also sets out three quantitative ceilings on long term debt. The first restricts debt service to 20 percent of revenues. The second limits the ratio of the stock of debt to revenues to 1.3:l. These restrictions are complementary. The first limits the proportion of total resources that can be used for debt service, but may be distorted by temporary reductions in debt service arising from grace periods and variable interest rates. These are captured in the debt stock: revenue ratio. "Revenues", in the case of long term debt, is defined as local taxes and fees, unconditional transfers and shared taxes-a somewhat more liberal definition than applies to short term debt. 24~awNo. 9869, Dt. 4.2.2008"On Local Borrowing" 100. The third quantitative ceiling requires an operating surplus coverage ratio (the ratio of operating surpluses to debt service) of at least 1.4:1. For purposes of this calculation, the operating surplus is defined as the difference between: (1) revenues from local taxes and fees, shared taxes and unconditional transfers and (2) unconditional operating expenditures25.This is a useful supplementary restriction on local borrowing. The operating surplus indicates the level of a local government's discretionary resources--i.e., revenues not pledged to wages, pension contributions, recurrent services and other non- interest recurrent obligations-that would be available to service debt in the event of an unexpected drop in revenues or increase in other recurrent expenditures. The specific ratio adopted by the Government is fairly liberal. In effect, the 1.4:1 ratio requires a local government to have financial cushion equal to 40 percent of its debt service obligations. Under these circumstances, a highly leveraged local government could find itself unable to service its debt if it experienced, for example, a relatively minor drop in revenues.26For this reason, the ceilings should be revisited in the light of actual experience after a period of, say, five years. Macroeconomic Limitations on Aggregate Debt 101. The rules for allocating available "debt space" between central and local government remain to be decided. In addition to the above limits, Article 20 allows the Government to fix the aggregate level of local government borrowing in order to ensure compliance with macro economic targets. The aggregate ceiling on total public sector borrowing is now set at 60 percent of GDP. Under the borrowing law, Parliament may determine the proportion of this amount that is available to local governments. This is to be set annually via the annual budget law. So far, neither the borrowing law nor the organic budget law defines how the Government will implement this provision: how the available 'debt space'--the difference between the existing debt stock and the level permitted by law--will be divided between the central government and the local governments as a group and how the local governments' share would be divided among individual jurisdictions. 102. For the present, this may not be an urgent question. The current stock of public sector debt is equivalent to 55 percent of GDP. If all local governments borrowed to the maximum permitted by the three quantitative restrictions in the borrowing law, their total stock would equal 2.8 percent of GDP- leaving room for the central government to grant all requests while continuing to add to its own stock of debt.27 If and when the central government debt does approach the ceiling, however, it will have to address this issue. One solution would be to authorize new borrowing to otherwise qualified local governments on a first-come, first-served basis, with the proviso that the disbursement of loan-tranches for ongoing construction projects receive priority over new starts. 250peratingexpenditures exclude interest payments. The term 'unconditional' excludes spending on functions financed through earmarked grants, such as the ndihma ekonomike,and from competitive grants. 2 6 ~ oexample, an LG with revenues of 100 million lek and current expenditures of 75 million lek would have an r operating surplus of 25 million lek, permitting debt service of 18 million lek. A ten percent drop in revenues would reduce the operating surplus to 15 million lek, which would be insufficient to cover existing debt service, even if all non-recurrent expendituresother than principal repayments were eliminated. 27Due to data constraints, this calculation does not reflect the impact of the ceiling on the operating surp1us:debt service ratio. The calculation is based on 2007 revenue and expenditure data for 373 local governments and assumes all loans are made at an interest rate of 10 percent with a maturity of ten years. Lower interest rates and longer maturities would, of course, increase the level of debt permitted by the three quantitativerestrictions. Fiscal Distress 103. The law also provides for measures to be taken if debt, once contracted, is not serviced. Article 23 of the law specifies a series of action actions to be taken in the event of fiscal distress. This is defined as a situation in which: (1) the level of accounts payable (including debt service) over 90 days in arrears exceeds 30 percent of the local government budget for that quarter or (2) any short term loans remain outstanding after November 30 of the current budget year. Under either of these conditions, a local government is forbidden to take any decision that 'could result in an increase in financial liabilities', introduce any new public services, or borrow without the approval of the MOF. The local government must also prepare a financial rehabilitation plan (with assistance of the MOF) and submit regular reports on its implementation to the MOF. The distress status ends only after the local government has run a budget surplus (or at least avoided a deficit) for two years in a row and the MOF finds that the local government has 'taken actions sufficient to eliminate the original source of problem.' 104. The law also makes a provision for a more severe financial condition, termed insolvency. This is defined as a situation in which (1) the level of contractual debt service more than 60 days in arrears exceeds 30 percent of total discretionary revenues or (2) the level of excessive accounts payable (as defined in the preceding article) has had notable negative repercussions for the delivery of basic services. The law specifies that the remedies to be applied in case of insolvency will be specified in separate legislation. The wording of an earlier version of the law would permit the courts to appoint an administrator with the power to approve all local expenditures, manage local government personnel (including hiring and firing of staff) and impose additional taxes, along with the power to impose a debt rescheduling plan, dismiss elected or appointed officials, and suspend labor agreements. 105. There are advantages and disadvantages to such provisions in a local borrowing law. On the positive side, they help central governments address a fundamental dilemma. Unlike private borrowers, local governments cannot be permitted to go out of business. They have to continue to provide key public services. At the same time, central government efforts to keep municipalities in business can create a moral hazard. Bailouts are likely to encourage local governments to engage in reckless behavior in the future. They are also likely to encourage lenders-including suppliers-to extend credit where it is not warranted by the financial condition of the borrower. In theory, fiscal distress measures provide an orderly mechanism to permit central governments to resolve the fiscal problems of distressed municipalities while imposing an appropriate degree of punishment for reckless fiscal behavior on both the local government and its creditors. 106. But there is negative side to such provisions. They can open the door to politically-motivated central government involvement, e.g., in deciding whether a particular local government decision 'could result in an increase in financial liabilities' or whether a local government has 'taken actions sufficient to eliminate the original source of problem.' 107. In the short term, the issue may be moot. It is not clear that the financial sector is prepared to play a major role in local infrastructure financing any time soon. It is true that the Albanian banking system has made remarkable progress since the 1997 crisis. State owned banks have been gradually privatized and there has been an increase in the number of private banks (from four in 1997 to 17 in 2006).28 ASa result of enhanced confidence in the domestic banking system, the ratio of banking assets to GDP has increased to 70 per cent. But lending to municipalities for infrastructure projects presents 28~ddressby Mr. Fatos Ibrahimi, Deputy Governor of the Bank of Albania, at the European Finance Convention, Tirana, 14-15 September 2007. particular risks. The physical assets of local governments make poor collateral: streets and schools cannot be sold to satisfy creditors. Lenders may also be justifiably leery of making loans that extend beyond the tenure of the existing mayor and council, fearing that winning candidates would renege on the obligations incurred by their predecessors. Box 5: Controlling Arrears Despite the provisions of the borrowing law, there is some risk that local governments could run up excessive levels of arrears to suppliers and contractors.Although such debts are, in principle, subject to the debt ceilings, the ceilings may be difficult to enforce in such cases. One might expect suppliers and contractors to voluntarily curtail credit if they are not paid. But experience suggests that such creditors may knowingly extend excessive credit to local governments if they expect the central government will eventually come to their relief. These expectations are often satisfied. Such relief, once granted, fosters the expectation of bailouts to follow encouraging further reckless lending. The Government should therefore avoid providing any financial relief to local governments that find themselves unable to pay their arrears. Local government and their creditors should instead be expected to work out payment arrangements bilaterally or through the courts. 108. In principle, local governments could provide some reassurance to lenders by pledging revenues. The borrowing law gives local governments the authority to pledge specific revenues for debt service and to authorize the Treasury to pay debt service directly to creditors by deducting it from unconditional transfers. But this carries certain risks. With debt service guaranteed by a Treasury intercept, lenders may be lax in assessing the debt servicing capacity of local government and may extend more credit than local governments are able to repay. Treasury should therefore permit such intercepts only on a selective basis. 109. Overall, in the light of these supply-side constraints, the Albanian government is likely to be in the business of mobilizing and allocating funds for local capital investment for some time to come. Priority should be given to improving the allocation system rather than to facilitating access to the private capital market. CHANGES THE STRUCTURE LOCALGOVERNMENT IN OF 110. The Government's decentralization strategy proposes fairly radical changes in the structure of subnational government, but provides no timetable for implementing them. Although this suggests that they are not regarded as urgent, they bear analysis. Regional Governments 111. Albania's new (1998) constitution established a new tier of subnational government: the regions. Unlike the prefectures and (now abolished) districts, these are not deconcentrated arms of the central government, but rather councils of local governments, whose membership consists of the mayors of the local governments within their territory and additional representatives from local government councils, with the number of seats allocated in proportion to each jurisdiction's population. The regional councils receive a small share of the unconditional transfer but have no independent taxing power. Their functional responsibilities are ill-defined--a point that is reiterated in the retrospective section of the strategy, although it notes that 'some adjustments have been made in compliance with law 8652 and policy documents in the areas of pre-university education, primary health care, social assistance, and social services.' 112. The decentralization strategy calls for transforming the regional councils into a second level of local government. This would involve electoral reforms (presumably including the direct election of regional councilors), changes in boundaries, and a clear definition of functional responsibilities and revenue sources. Legislation to this effect is to be enacted by January 2009. The strategy cites several possible functional roles for regional governments, including water supply and sewerage, regional roads, solid waste collection and disposal, urban planning, and regional social, cultural, and economic activities. In principle, these roles are reasonable. They are all functions where some degree of inter-municipal collaboration may be useful: bulk water supply storage and conveyance, regional sewage treatment, regional solid waste disposal, inter-municipal roads, and the coordination of urban plans and regional cultural activities. 113. But the notion that a country the size of Albania needs a third tier of government to perform these roles is questionable. With respect to bulk water supply, regional water companies already exist. In fact, the Government is taking pains to keep them intact even as it transfers their ownership to the local governments. In the roads sector, a regional tier of government would be superfluous if, as proposed in connection with the recently approved Secondary and Local Roads Project, regional access roads are transferred to the central government. Regional solid waste disposal can be managed by a special purpose entity. 114. It should also be noted that Albania need not establish regional governments to access EU structural funds. Under the regulations governing EU pre-accession funds for the 2006-2013 planning period, Albania is not eligible to receive funds under the IPAYsregional development component.29 If Albania were to be declared eligible, it would not, in any event, be required to establish or maintain regional units of government. The current regulations governing the regional development component require only that each operational program financed by the component 'include a justification for the thematic, geographical and financial concentration on its priorities' and information on the procedures the Government expects to use to allocate and monitor these funds.30The Government would therefore be well advised to investigate other, less cumbersome, methods of solving specific inter-municipal coordination problems before creating full-fledged regional governments. Consolidation 115. The decentralization strategy also calls for the consolidation of small communes, citing their generally low level of technical competence and their inefficient use of financial resources (particularly their high administrative overheads). As evidence, it states that in 80 communes--mainly those with populations below 5,000--more than 70 percent of staff has no higher education and that 220 communes spend over 40 percent of their budgets on administration. Whether a higher education is required to perform the duties of most staff in communes is subject to debate. The significance of high levels of administrative spending is also questionable. As noted earlier, smaller municipalities are said to classify all spending on personnel as administrative. 116. Whether the figures are correct or not, the strategy's conclusions are debatable. As shown in the table on the first page of this report, Albanian local governments are not particularly small by 29~uropeanUnion. Council RegulationNo. 108512006, 30 European Union. Council Regulation No. 108312006, European standards. Moreover, international efforts to demonstrate economies of scale in the provision of municipal services have generally proven inconcl~sive.~~ 117. Overall, the evidence suggests that population size is not a decisive variable in determining the cost or quality of public services. The literature suggests, in fact, that where populations are geographically dispersed, there are few economies of scale to be gained by incorporating them into a single large jurisdiction. International experience also suggests that difficulties in attracting competent staff can be addressed by contracting with private providers, joint service arrangements, or by employing skilled staff on a part time basis. 118. There is a final lesson to be drawn from international experience: territorial consolidation can be extremely unpopular. Small communities resist incorporation into larger jurisdictions, fearing their interests will be submerged in the larger polity. Among the recent EU accession countries, only one (Lithuania) has successfully implemented a territorial consolidation in the post-Soviet period. If such resistance exists in Albania, the costs of attempting consolidation may not be worth the effort. Efforts at improving the efficiency of local government could instead focus on other options. One is the use of the private contractors, who can achieve economies of scale by providing services to several different jurisdictions at the same time. The small communes of France, for example, rely on large private water companies to provide drinking water. Another is the use of inter-municipal service companies. Small municipalities in Latvia rely on jointly owned enterprises to operate joint land fill sites. These have a proven track record and bear consideration in Albania. 3 ' ~ h iliterature is surveyed in Dillinger, William. 2003. Latvia: Beyond Territorial Reform (World Bank Report s No. 25466-LV) and in William F. Fox and Tami Gurley. 2005. Will Consolidation Improve Sub-National Governments?